目录
一、转换器与估计器
二、分类算法
K-近邻算法
案例代码:
模型选择与调优
案例代码:
朴素贝叶斯算法:
朴素贝叶斯算法总结
案例代码:
决策树总结:
案例代码:
使用随机森林来实现:
随机森林总结
总结
本次案例的代码集:
一、转换器与估计器 二、分类算法 K-近邻算法
K-近邻算法
KNN算法总结:
优点:
简单、易于理解、易于实现、无需训练
缺点:
1)必须指定K值,K值选定不当则分类精度不能保证。
2)懒惰算法,对测试样本分类时的计算量大,内存开销大
使用场景:
小数据场景,几千~几万条样本,具体使用看业务场景。
案例代码:
from sklearn.datasets import load_irisfrom sklearn.model_selection import train_test_splitfrom sklearn.preprocessing import StandardScalerfrom sklearn.neighbors import KNeighborsClassifierdef knn_iris(): """ 用KNN算法对iris数据进行分类 :return: """ # 1)获取数据 iris = load_iris() # 2)划分数据集 x_train, x_test, y_train, y_test = train_test_split(iris.data, iris.target, random_state=6 ) # 3) 特征工程:标准化 transfer = StandardScaler() x_train = transfer.fit_transform(x_train) x_test = transfer.transform(x_test) # 4) KNN算法预估器 estimator = KNeighborsClassifier(n_neighbors=3) estimator.fit(x_train, y_train) # 5) 模型评估 # 方法1:直接比对真实值和预测值 y_predict = estimator.predict(x_test) print("y_predict:n", y_predict) print("直接比对真实值和预测值:n", y_test == y_predict) # 方法2: 计算准确率 score = estimator.score(x_test, y_test) print("准确率为:n", score) return Noneif __name__ == '__main__': # 代码1:用KNN算法对iris数据进行分类 knn_iris()
模型选择与调优
案例代码
from sklearn.datasets import load_irisfrom sklearn.model_selection import train_test_splitfrom sklearn.preprocessing import StandardScalerfrom sklearn.neighbors import KNeighborsClassifierfrom sklearn.model_selection import GridSearchCVdef knn_iris_gscv(): """ 用KNN算法对iris数据进行分类,添加网格搜索和交叉验证 :return: """ # 1)获取数据 iris = load_iris() # 2)划分数据集 x_train, x_test, y_train, y_test = train_test_split(iris.data, iris.target, random_state=6 ) # 3) 特征工程:标准化 transfer = StandardScaler() x_train = transfer.fit_transform(x_train) x_test = transfer.transform(x_test) # 4) KNN算法预估器 estimator = KNeighborsClassifier() # 加入网格搜索和交叉验证 # 参数准备 param_dict = {"n_neighbors": [1, 3, 5, 7, 9, 11]} estimator = GridSearchCV(estimator, param_grid=param_dict, cv=10) estimator.fit(x_train, y_train) # 5) 模型评估 # 方法1:直接比对真实值和预测值 y_predict = estimator.predict(x_test) print("y_predict:n", y_predict) print("直接比对真实值和预测值:n", y_test == y_predict) # 方法2: 计算准确率 score = estimator.score(x_test, y_test) print("准确率为:n", score) # 最佳参数结果:best_param_ print("最佳参数:n", estimator.best_params_) # 最佳结果:best_score_ print("最佳结果:n", estimator.best_score_) # 最佳估计器:best_estimator_ print("最佳估计器:n", estimator.best_estimator_) # 交叉验证结果: cv_results_ print("交叉验证结果:n", estimator.cv_results_) return Noneif __name__ == '__main__': # 代码2: 用KNN算法对iris数据进行分类,添加网格搜索和交叉验证 knn_iris_gscv()
facebook数据挖掘案例:
案例代码:import pandas as pdfrom sklearn.model_selection import train_test_split, GridSearchCVfrom sklearn.neighbors import KNeighborsClassifierfrom sklearn.preprocessing import StandardScalerdef predict_data(): """ 数据预处理 :return: """ # 1)读取数据 data = pd.read_csv("./train.csv") # 2)基本数据处理 # 缩小范围 data = data.query("x<2.5 & x>2 & y<1.5 & y>1.0") # 处理时间特征 time_value = pd.to_datatime(data["time"], unit="s") date = pd.DatetimeIndex(time_value) data.loc[:, "day"] = date.day data.loc[:, "weekday"] = date.weekday data["hour"] = data.hour # 3)过滤签到次数少的地点 data.groupby("place_id").count() place_count = data.groupby("place_id").count()["row_id"] data_final = data[data['place_id'].isin(place_count[place_count > 3].index.vlaues)] # 筛选特征值和目标值 x = data_final[["x", "y", "accuracy", "day", "weekday", "hour"]] y = data_final["place_id"] # 数据集划分 # 机器学习 x_train, x_test, y_train, y_test = train_test_split(x, y) # 3) 特征工程:标准化 transfer = StandardScaler() x_train = transfer.fit_transform(x_train) x_test = transfer.transform(x_test) # 4) KNN算法预估器 estimator = KNeighborsClassifier() # 加入网格搜索和交叉验证 # 参数准备 param_dict = {"n_neighbors": [1, 3, 5, 7, 9, 11]} estimator = GridSearchCV(estimator, param_grid=param_dict, cv=3) estimator.fit(x_train, y_train) # 5) 模型评估 # 方法1:直接比对真实值和预测值 y_predict = estimator.predict(x_test) print("y_predict:n", y_predict) print("直接比对真实值和预测值:n", y_test == y_predict) # 方法2: 计算准确率 score = estimator.score(x_test, y_test) print("准确率为:n", score) # 最佳参数结果:best_param_ print("最佳参数:n", estimator.best_params_) # 最佳结果:best_score_ print("最佳结果:n", estimator.best_score_) # 最佳估计器:best_estimator_ print("最佳估计器:n", estimator.best_estimator_) # 交叉验证结果: cv_results_ print("交叉验证结果:n", estimator.cv_results_) return Noneif __name__ == '__main__': predict_data()
朴素贝叶斯算法:
案例代码
from sklearn.datasets import load_irisfrom sklearn.model_selection import train_test_splitfrom sklearn.preprocessing import StandardScalerfrom sklearn.neighbors import KNeighborsClassifierfrom sklearn.model_selection import GridSearchCVfrom sklearn.datasets import fetch_20newsgroupsfrom sklearn.feature_extraction.text import TfidfVectorizerfrom sklearn.naive_bayes import MultinomialNBdef nb_news(): """ 用朴素贝叶斯算法对新闻进行分类 :return: """ # 1)获取数据 news = fetch_20newsgroups(subset="all") # 2)划分数据集 x_train, x_test, y_train, y_test = train_test_split(news.data, news.target) # 3)特征工程文本特征抽取-tfidf transfer = TfidfVectorizer() x_train = transfer.fit_transform(x_train) x_test = transfer.transform(x_test) # 4)朴素贝叶斯算法预估器流程 estimator = MultinomialNB() estimator.fit(x_train, y_train) # 5)模型评估 # 方法1:直接比对真实值和预测值 y_predict = estimator.predict(x_test) print("y_predict:n", y_predict) print("直接比对真实值和预测值:n", y_test == y_predict) # 方法2:计算准确率 score = estimator.score(x_test, y_test) print("准确率为:n", score) return Noneif __name__ == '__main__': # 代码3:用朴素贝叶斯算法对新闻进行分类 nb_news()
朴素贝叶斯算法总结
优点:
对缺失数据不太敏感,算法比较简单,常用于文本分类。
分类准确度高,速度快。
缺点:
由于使用样本独立的假设,所以如果特征之间关联,预测效果不明显。
决策树
案例代码:from sklearn.datasets import load_irisfrom sklearn.model_selection import train_test_splitfrom sklearn.preprocessing import StandardScalerfrom sklearn.neighbors import KNeighborsClassifierfrom sklearn.model_selection import GridSearchCVfrom sklearn.datasets import fetch_20newsgroupsfrom sklearn.feature_extraction.text import TfidfVectorizerfrom sklearn.naive_bayes import MultinomialNBfrom sklearn.tree import DecisionTreeClassifier, export_graphvizdef decision_iris(): """ 用决策树对iris数据进行分类 :return: """ # 1)获取数据集 iris = load_iris() # 2)划分数据集 x_train, x_test, y_train, y_test = train_test_split(iris.data, iris.target, random_state=22) # 3)决策树预估器 estimator = DecisionTreeClassifier(criterion="entropy") estimator.fit(x_train, y_train) # 4)模型评估 # 方法1:直接比对真实值和预测值 y_predict = estimator.predict(x_test) print("y_predict:n", y_predict) print("直接比对真实值和预测值:n", y_test == y_predict) # 方法2: 计算准确率 score = estimator.score(x_test, y_test) print("准确率为:n", score) # 可视化决策树 export_graphviz(estimator, out_file="iris_tree.dot", feature_names=iris.feature_names) return Noneif __name__ == '__main__': # 代码4:用决策树对iris数据进行分类 decision_iris()
决策树支持可视化:
.dot文件转换为可视化图像的网页:Graphviz Online
决策树总结:
优点:
可视化——解释性强
缺点:
容易产生过拟合,这时候使用随机森林效果会好些
决策树的实验项目——titanic数据的案例
案例代码:import pandas as pdfrom sklearn.feature_extraction import DictVectorizerfrom sklearn.model_selection import train_test_splitfrom sklearn.tree import DecisionTreeClassifier, export_graphvizdef decision_titanic(): # 1、获取数据 titanic = pd.read_csv("./titanic.csv") print(titanic) # 筛选特征值和目标值 x = titanic[["pclass", "age", "sex"]] y = titanic["survived"] # 2、数据处理 # 1)缺失值处理 x['age'].fillna(x["age"].mean(), inplace=True) # 2)转换成字典 x = x.to_dict(orient="records") # 3、数据集划分 x_train, x_test, y_train, y_test = train_test_split(x, y, random_state=22) transfer = DictVectorizer() x_train = transfer.fit_transform(x_train) x_test = transfer.transform(x_test) # 3)决策树预估器 estimator = DecisionTreeClassifier(criterion="entropy", max_depth=8) estimator.fit(x_train, y_train) # 4)模型评估 # 方法1:直接比对真实值和预测值 y_predict = estimator.predict(x_test) print("y_predict:n", y_predict) print("直接比对真实值和预测值:n", y_test == y_predict) # 方法2:计算准确率 score = estimator.score(x_test, y_test) print("准确率为:n", score) # 可视化决策树 export_graphviz(estimator, out_file="titanic_tree.dot", feature_names=transfer.get_feature_names())if __name__ == '__main__': decision_titanic()
使用随机森林来实现:
import pandas as pdfrom sklearn.feature_extraction import DictVectorizerfrom sklearn.model_selection import train_test_splitfrom sklearn.tree import DecisionTreeClassifier, export_graphvizfrom sklearn.ensemble import RandomForestClassifierfrom sklearn.model_selection import GridSearchCVdef decision_titanic(): # 1、获取数据 titanic = pd.read_csv("./titanic.csv") print(titanic) # 筛选特征值和目标值 x = titanic[["pclass", "age", "sex"]] y = titanic["survived"] # 2、数据处理 # 1)缺失值处理 x['age'].fillna(x["age"].mean(), inplace=True) # 2)转换成字典 x = x.to_dict(orient="records") # 3、数据集划分 x_train, x_test, y_train, y_test = train_test_split(x, y, random_state=22) transfer = DictVectorizer() x_train = transfer.fit_transform(x_train) x_test = transfer.transform(x_test) # 3)随机森林预估器 estimator = RandomForestClassifier() # 加入网格搜索与交叉验证 # 参数准备 param_dict = {"n_estimators": [120, 200, 300, 500, 800, 1200], "max_depth": [5, 8, 15, 25, 30]} estimator = GridSearchCV(estimator, param_grid=param_dict, cv=3) estimator.fit(x_train, y_train) # 4)模型评估 # 方法1:直接比对真实值和预测值 y_predict = estimator.predict(x_test) print("y_predict:n", y_predict) print("直接比对真实值和预测值:n", y_test == y_predict) # 方法2:计算准确率 score = estimator.score(x_test, y_test) print("准确率为:n", score) # 可视化决策树 export_graphviz(estimator, out_file="titanic_tree.dot", feature_names=transfer.get_feature_names())if __name__ == '__main__': decision_titanic()
随机森林总结
优点:
能够有效的运行在大数据集上
处理具有高维特征的输入样本,而且不需要降维。
总结
本次案例的代码集:
from sklearn.datasets import load_irisfrom sklearn.model_selection import train_test_splitfrom sklearn.preprocessing import StandardScalerfrom sklearn.neighbors import KNeighborsClassifierfrom sklearn.model_selection import GridSearchCVfrom sklearn.datasets import fetch_20newsgroupsfrom sklearn.feature_extraction.text import TfidfVectorizerfrom sklearn.naive_bayes import MultinomialNBfrom sklearn.tree import DecisionTreeClassifier, export_graphvizdef knn_iris(): """ 用KNN算法对iris数据进行分类 :return: """ # 1)获取数据 iris = load_iris() # 2)划分数据集 x_train, x_test, y_train, y_test = train_test_split(iris.data, iris.target, random_state=6 ) # 3) 特征工程:标准化 transfer = StandardScaler() x_train = transfer.fit_transform(x_train) x_test = transfer.transform(x_test) # 4) KNN算法预估器 estimator = KNeighborsClassifier(n_neighbors=3) estimator.fit(x_train, y_train) # 5) 模型评估 # 方法1:直接比对真实值和预测值 y_predict = estimator.predict(x_test) print("y_predict:n", y_predict) print("直接比对真实值和预测值:n", y_test == y_predict) # 方法2: 计算准确率 score = estimator.score(x_test, y_test) print("准确率为:n", score) return Nonedef knn_iris_gscv(): """ 用KNN算法对iris数据进行分类,添加网格搜索和交叉验证 :return: """ # 1)获取数据 iris = load_iris() # 2)划分数据集 x_train, x_test, y_train, y_test = train_test_split(iris.data, iris.target, random_state=6 ) # 3) 特征工程:标准化 transfer = StandardScaler() x_train = transfer.fit_transform(x_train) x_test = transfer.transform(x_test) # 4) KNN算法预估器 estimator = KNeighborsClassifier() # 加入网格搜索和交叉验证 # 参数准备 param_dict = {"n_neighbors": [1, 3, 5, 7, 9, 11]} estimator = GridSearchCV(estimator, param_grid=param_dict, cv=10) estimator.fit(x_train, y_train) # 5) 模型评估 # 方法1:直接比对真实值和预测值 y_predict = estimator.predict(x_test) print("y_predict:n", y_predict) print("直接比对真实值和预测值:n", y_test == y_predict) # 方法2: 计算准确率 score = estimator.score(x_test, y_test) print("准确率为:n", score) # 最佳参数结果:best_param_ print("最佳参数:n", estimator.best_params_) # 最佳结果:best_score_ print("最佳结果:n", estimator.best_score_) # 最佳估计器:best_estimator_ print("最佳估计器:n", estimator.best_estimator_) # 交叉验证结果: cv_results_ print("交叉验证结果:n", estimator.cv_results_) return Nonedef nb_news(): """ 用朴素贝叶斯算法对新闻进行分类 :return: """ # 1)获取数据 news = fetch_20newsgroups(subset="all") # 2)划分数据集 x_train, x_test, y_train, y_test = train_test_split(news.data, news.target) # 3)特征工程文本特征抽取-tfidf transfer = TfidfVectorizer() x_train = transfer.fit_transform(x_train) x_test = transfer.transform(x_test) # 4)朴素贝叶斯算法预估器流程 estimator = MultinomialNB() estimator.fit(x_train, y_train) # 5)模型评估 # 方法1:直接比对真实值和预测值 y_predict = estimator.predict(x_test) print("y_predict:n", y_predict) print("直接比对真实值和预测值:n", y_test == y_predict) # 方法2:计算准确率 score = estimator.score(x_test, y_test) print("准确率为:n", score) return Nonedef decision_iris(): """ 用决策树对iris数据进行分类 :return: """ # 1)获取数据集 iris = load_iris() # 2)划分数据集 x_train, x_test, y_train, y_test = train_test_split(iris.data, iris.target, random_state=22) # 3)决策树预估器 estimator = DecisionTreeClassifier(criterion="entropy") estimator.fit(x_train, y_train) # 4)模型评估 # 方法1:直接比对真实值和预测值 y_predict = estimator.predict(x_test) print("y_predict:n", y_predict) print("直接比对真实值和预测值:n", y_test == y_predict) # 方法2: 计算准确率 score = estimator.score(x_test, y_test) print("准确率为:n", score) # 可视化决策树 export_graphviz(estimator, out_file="iris_tree.dot", feature_names=iris.feature_names) return Noneif __name__ == '__main__': # 代码1:用KNN算法对iris数据进行分类 # knn_iris() # 代码2: 用KNN算法对iris数据进行分类,添加网格搜索和交叉验证 # knn_iris_gscv() # 代码3:用朴素贝叶斯算法对新闻进行分类 # nb_news() # 代码4:用决策树对iris数据进行分类 decision_iris()